1 / 16

Crowdsourcing & split-tunneling to map and circumvent URL filtering

Crowdsourcing & split-tunneling to map and circumvent URL filtering. Walid Al- Saqaf Örebro University - Sweden A presentation for RIPE 68 Warsaw, May 14. Two parallel tracks using Alkasir. Mapping URL filtering through crowdsourcing - Reporting of suspected filtered URLs

cheng
Download Presentation

Crowdsourcing & split-tunneling to map and circumvent URL filtering

An Image/Link below is provided (as is) to download presentation Download Policy: Content on the Website is provided to you AS IS for your information and personal use and may not be sold / licensed / shared on other websites without getting consent from its author. Content is provided to you AS IS for your information and personal use only. Download presentation by click this link. While downloading, if for some reason you are not able to download a presentation, the publisher may have deleted the file from their server. During download, if you can't get a presentation, the file might be deleted by the publisher.

E N D

Presentation Transcript


  1. Crowdsourcing & split-tunneling to map and circumvent URL filtering Walid Al-Saqaf Örebro University - Sweden A presentation for RIPE 68 Warsaw, May 14

  2. Two parallel tracks using Alkasir Mapping URL filtering through crowdsourcing - Reporting of suspected filtered URLs Manual submissions & automatic cross-validation with crowdsourcing - Who is blocking? Autonomous Systems or Internet Service Providers? Crowdsourcing to identify similar response patterns to HTTP requests 2) Circumventing filtering through SSH split-tunneling Develop an effective and affordable circumvention method Users to circumvent and also provide data Analysis of user behavior in reporting about censorship

  3. 1) Mapping URL Filtering Manual entry of URL (one time process) If filtering identified (to be explained in the next slide) Automatic crowdsourcing triggered for other users in the same country Content analysis using statistical software (SPSS)

  4. DEMO User reports a URL manually Software verifies –using server- by comparing headers and/or content with remotely fetched version of the same URL If found to be blocked, the moderator verifies if the URL is not of porn/nudity content (if approved before, it is approved automatically & vice versa) Moderator is a closed group and public access is limited due to personal risk Assessment of what is porn and what is not is based on meta website data Entry is added to the database with relevant data accordingly Requests to clients of other users using the same AS/ISP are sent to also check whether the URL is filtered (all in the background)* If multiple clients confirmed the filtering, the request is sent to clients of other users in the same country but using other ISPs* Note: It is also possible to check manually entered URLs in any country/ISP in background * Under development URL Submissions

  5. Different HTTP response status code (403, 504…) (exception: redirections to localized versions or forbidden by the source due to geo-location …) No header response (connectiondropped) How to know if a URL is filtered (assuming right IP address reached, i.e., no DNS tampering occurred) ? Substantial modified content check the first 1000k and compare text (not bullet-proof) 200 but with different HTTP header response (e.g., common firewall signatures [Websense, SmartFilter, etc.], significantly different content size

  6. Example: Bahrain, same ISP No consistency & no transparency

  7. Maximizes productivity by giving users the chance to decide what filtered URLs they want to access Enhances reliability through cross-validation based on location Allows testing and verifying any URL virtually any time and anywhere an Alkasir client is running When enabled, can track the level and depth of filtering through automatic addition of externally linked URLs in a reported document Benefits of source-crowding

  8. Socks local proxy activated after connecting to SSH server (SSL and obfuscated/normal Putty) with local dynamic port (AES CBC enc for SSH2) A pool of 100+ proxy IPs available (different class B addresses) Proxy automatic configuration file that is retrieved when Alkasir is connected is used to divert traffic as required (split-tunneling) Remote DNS resolution enabled to prevent DNS leaks Alkasir.com is always accessed through the encrypted tunnel 2) Circumventing URL Filtering Blank alkasir.pac file: function FindProxyForURL(url,host) { if (shExpMatch(host, "*alkasir.com*")) { return "SOCKS5 127.0.0.1:<local PORT>; DIRECT"; } return "DIRECT"; }

  9. Alkasir user Encrypted & tunneled (hidden from ISP) Regular (exposed to ISP) UNCENSORED The split-tunneling approach The PAC file is built based on the geo-location of the client retrieved through ip2location (paid service) with the IP as input

  10. Utilizes bandwidth on the server and client efficiently Allows users to access localized services, news normally Reduces the risk of proxy-use detection by the ISP Encourages reporting more filtered URLs, moderators to learn more about filtering methods and content Can serve as an additional plug-in for other circumvention tools (utilizing the list of blocked URLs) Encourages the updating of empirical data on filtering for use by advocacy groups, scholars, and users Benefits of split-tunneling

  11. Deriving the sample (1/10/2010-1/10/2012)

  12. Server data shows: Filtering patterns corresponded to political developments A combined graph showing the number of URL reports for the four case studies (Tunisia, Egypt, Yemen, and Syria)

  13. Survey: Most widely used circumvention tools

  14. Survey: What is needed for better circumvention

  15. Enhancing Alkasir’s blocking-resistance • Developing extensions for browsers and apps for mobile devices • Reducing risks to users, ensuring lack of personal data and enhancing split-tunneling to prevent detection of proxy use • Adding multiple circumvention methods (e.g., VPN, web-based…) • Building an API to allow access to data (requires risk assessment) • Defining a more concrete & systematic categorization approach • Cooperating with like-minded projects to improve reliability of data (Herdict, OpenNet Initiative, Choke Point) • Proposing the integration of Alkasir’s data into other popular circumvention tools (Tor, Psiphon, etc.) to enable split-tunneling Future work and ambitions

  16. Questions

More Related